Machine Learning for Hybrid & Electric Vehicle Battery Prognostics

The goal of this project is to decrease battery reconditioning and remanufacturing time by expanding the number of predictor variables from existing reconditioning equipment for use in exponential gaussian process regression machine learning models to shorten the time required for determination of amp-hour capacities and for categorizing modules for the aftermarket, reuse or crush-and-shred (recycling).

This project is expected to reduce primary feedstock by 0.015 million metric tons (MMT) of EV batteries retiring in US (2030), and reduce energy consumption by 3.6 PJ.